478 ALASKAN GLACIER STUDIES harmonizes with the theory that the Prince William Sound Glacier spread out into a great bulb on emerging into the sea and that it found freest passage and moved fastest in the narrow strait between Hinchinbrook and Montague Island and accordingly scoured out the passage through which it flowed. It could, of course, have produced the great depth in Hinchinbrook Entrance without extending beyond the islands, provided the ice moved through this passage long enough, or in a shorter period if the ice moved some distance out to sea. It is not absolutely clear, however, that this has taken place and that the continental shelf is glaciated. OUT doubt is based, chiefly, upon the low limit of glaciation at Hinchinbrook Island. If the ice extended only 400 feet above sea level and sloped between 25 and 60 feet to the mile, as in Prince William Sound, it could not well have advanced fifty-six miles farther to Middleton Island. Another objection is found in the fact that the nature of the material from Middleton Island, which Dawson interprets as till, is such that it might equally well be an uplifted marine clay or silt, sprinkled with sub-angular, striated stones dropped by floating icebergs from glacier fronts in Hinchinbrook Entrance and the other outlets of Prince William Sound. The presence of crushed shells in the Middleton Island deposit and the fact of known, post-glacial uplift a short distance to the east on Wingham Island* support this interpretation. The present conclusion therefore, regarding the glaciation of the broad continental shelf off Prince William Sound is that it is not certain whether the shelf has been actually overridden by an ice sheet or not, but that it has certainly been greatly modified through the former presence of glaciers, at least at the mouths of the various outlets of Prince William Sound. In any event, there are vast glacial accumulations on the continental shelf, made either by the deposition of till or of great volumes of glacial silt, supplied by rivers from beneath the melting ice, and distributed by the tidal currents, as at the Copper River delta today, with a scattering of glacial bowlders carried by icebergs. This silt-and-iceberg-bowlder combination is hard to distinguish from glacial till, or bowlder clay, unless marine shells are buried in it. The Old Rock Floor of Prince William Sound. In various parts of Prince William Sound there are rock terraces at elevations of from 40 to 100 or 200 feet above sea level. These were first noted by Schrader.8 In 1910 we saw others, which lead us to believe that these are remnants of the old rock floor of Prince William Sound. Our view is that there was no preglacial arm of the sea on the present site of Prince William Sound, but that glacial erosion has excavated the whole sound on the site of a preglacial lowland, whose rock floor was slightly above the present sea level. This lowland may have been crossed by preglacial stream courses, along the lines of which the fiords and the deeper submarine channels were excavated much more deeply than the intermediate areas. This view is supported by the character of the low, flat-topped, rock terraces observed. On Culross Island near Passage Canal, for example, there is a low terrace which disappears to the eastward near Prince William Sound and to the westward near Culross »Martin, G. C., Bull, 385, U. S. Geol. Survey, 1908, pp. 46, 64. »Schrader, F. C., 20th Ann. Kept., U. S. Geol. Survey, Part VJI, 1900, p. 404; Schrader and Spencer, House Doc. 546, 56th Congress, 2nd session, Washington, 1901, pp. 75-76. Gilbert, G. K., Harriman Alaska Expedition, Vol. HI, 1904, p. 176; Giant, U. S. and Higgins, D. F., Bull, 443, U. S. Geol. Survey, 1910, pp. 16-17.